The present invention relates generally to data communication, and specifically to a method, and apparatus implementing that method, for high-speed, synchronous transmission of data.
In data communications topologies, for example those used for local area networks, data is often communicated by a transmission medium comprising some form of physical signal path such as electrical wires, printed circuit paths and the like. Often, for high-speed, synchronous data transmissions and higher throughput, data bits are transmitted in parallel together with a clock signal. A drawback to this form of data transmission is that there always seems to be a need for one more data bit path when one does not exist. Thus, when there are less paths than data bit signals (i.e., a parallel grouping of data bits), resort must be made to a parallel/serial scheme of data transmission. For example, suppose data is desired to be synchronously communicated from a source to a destination in eight bit quantities, accompanied by a clock signal, but there are only eight signal paths available. One solution is to transmit, for each eight bit quantity, two groups of bits. But this degrades performance by approximately one-half.
Another solution is to design the system to use an encoding scheme whereby at least one data bit is encoded (such as Manchester encoding) to include the clock signal that is extracted at the destination. However, this requires that the underlying design incorporate this solution by including the necessary encoding circuitry. Another solution must be found if an existing design, such as off-the-shelf elements, is to be used.